Implement the first set of rounds of SHA-1 in assembly

2 files changed, 85 insertions, 5 deletions

diff --git a/modules/sha_x86/sha160.cpp b/modules/sha_x86/sha160.cpp
index be8b7f695..fce5c9d04 100644
--- a/modules/sha_x86/sha160.cpp
+++ b/modules/sha_x86/sha160.cpp
@@ -56,25 +56,41 @@ extern "C" void sha160_core(u32bit[5], const byte[64], u32bit[80]);
 void SHA_160::hash(const byte input[])
    {
 #if 1
-   sha160_core(digest, input, W);
+   u32bit digestX[5];
+   for(int j = 0; j != 5; j++)
+      digestX[j] = digest[j];
+
+   sha160_core(digestX, input, W);
+
+   u32bit A = digestX[0], B = digestX[1], C = digestX[2],
+          D = digestX[3], E = digestX[4];
+
 #else
    for(u32bit j = 0; j != 16; ++j)
       W[j] = make_u32bit(input[4*j], input[4*j+1], input[4*j+2], input[4*j+3]);
 
    for(u32bit j = 16; j != 80; ++j)
       W[j] = rotate_left((W[j-3] ^ W[j-8] ^ W[j-14] ^ W[j-16]), 1);
-#endif
 
    u32bit A = digest[0], B = digest[1], C = digest[2],
           D = digest[3], E = digest[4];
+#endif
 
-   F1(A,B,C,D,E,W[ 0]);   F1(E,A,B,C,D,W[ 1]);   F1(D,E,A,B,C,W[ 2]);
-   F1(C,D,E,A,B,W[ 3]);   F1(B,C,D,E,A,W[ 4]);   F1(A,B,C,D,E,W[ 5]);
+   /*
+   F1(A,B,C,D,E,W[ 0]); F1(E,A,B,C,D,W[ 1]);   F1(D,E,A,B,C,W[ 2]);
+   F1(C,D,E,A,B,W[ 3]);   F1(B,C,D,E,A,W[ 4]);
+
+
+   F1(A,B,C,D,E,W[ 5]);
    F1(E,A,B,C,D,W[ 6]);   F1(D,E,A,B,C,W[ 7]);   F1(C,D,E,A,B,W[ 8]);
-   F1(B,C,D,E,A,W[ 9]);   F1(A,B,C,D,E,W[10]);   F1(E,A,B,C,D,W[11]);
+   F1(B,C,D,E,A,W[ 9]);
+
+   F1(A,B,C,D,E,W[10]);   F1(E,A,B,C,D,W[11]);
    F1(D,E,A,B,C,W[12]);   F1(C,D,E,A,B,W[13]);   F1(B,C,D,E,A,W[14]);
+
    F1(A,B,C,D,E,W[15]);   F1(E,A,B,C,D,W[16]);   F1(D,E,A,B,C,W[17]);
    F1(C,D,E,A,B,W[18]);   F1(B,C,D,E,A,W[19]);
+   */
 
    F2(A,B,C,D,E,W[20]);   F2(E,A,B,C,D,W[21]);   F2(D,E,A,B,C,W[22]);
    F2(C,D,E,A,B,W[23]);   F2(B,C,D,E,A,W[24]);   F2(A,B,C,D,E,W[25]);
diff --git a/modules/sha_x86/sha1core.S b/modules/sha_x86/sha1core.S
index 38d8525db..5a09cb00d 100644
--- a/modules/sha_x86/sha1core.S
+++ b/modules/sha_x86/sha1core.S
@@ -28,6 +28,7 @@ sha160_core:
 
      leal    64(%edi), %edx
     
+        .p2align 4,,7
        // here esi == 16
 .EXPANSION_LOOP:
         movl    -32(%edx), %eax
@@ -42,6 +43,69 @@ sha160_core:
         cmpl    $80, %esi
         jne .EXPANSION_LOOP
 
+   // here: edi = W
+   
+   movl 20(%esp), %ebp
+   movl 0(%ebp), %eax
+   movl 4(%ebp), %ebx
+   movl 8(%ebp), %ecx
+   movl 12(%ebp), %edx
+   movl 16(%ebp), %esi
+
+#define MAGIC1 $0x5A827999
+
+#define FUNC1(B, C, D, TEMP)   \
+   movl C, TEMP                 ; \
+   xorl D, TEMP                 ; \
+   andl B, TEMP                 ; \
+   xorl D, TEMP
+
+#define FUNC2(B, C, D, TEMP)   \
+   movl B, TEMP                 ; \
+   xorl C, TEMP                 ; \
+   xorl D, TEMP  
+
+#define F(A, B, C, D, E, TEMP, MSG, MAGIC, FUNC) \
+   FUNC(B, C, D, TEMP)          ; \
+   addl MSG, E                  ; \
+   addl TEMP, E                 ; \
+   addl MAGIC, E                ; \
+   roll $5, A                   ; \
+   addl A, E                    ; \
+   rorl $5, A                   ; \
+   roll $30, B
+
+    F(%eax, %ebx, %ecx, %edx, %esi, %ebp,  0(%edi), MAGIC1, FUNC1)
+    F(%esi, %eax, %ebx, %ecx, %edx, %ebp,  4(%edi), MAGIC1, FUNC1)
+    F(%edx, %esi, %eax, %ebx, %ecx, %ebp,  8(%edi), MAGIC1, FUNC1)
+    F(%ecx, %edx, %esi, %eax, %ebx, %ebp, 12(%edi), MAGIC1, FUNC1)
+    F(%ebx, %ecx, %edx, %esi, %eax, %ebp, 16(%edi), MAGIC1, FUNC1)
+
+    F(%eax, %ebx, %ecx, %edx, %esi, %ebp, 20(%edi), MAGIC1, FUNC1)
+    F(%esi, %eax, %ebx, %ecx, %edx, %ebp, 24(%edi), MAGIC1, FUNC1)
+    F(%edx, %esi, %eax, %ebx, %ecx, %ebp, 28(%edi), MAGIC1, FUNC1)
+    F(%ecx, %edx, %esi, %eax, %ebx, %ebp, 32(%edi), MAGIC1, FUNC1)
+    F(%ebx, %ecx, %edx, %esi, %eax, %ebp, 36(%edi), MAGIC1, FUNC1)
+
+    F(%eax, %ebx, %ecx, %edx, %esi, %ebp, 40(%edi), MAGIC1, FUNC1)
+    F(%esi, %eax, %ebx, %ecx, %edx, %ebp, 44(%edi), MAGIC1, FUNC1)
+    F(%edx, %esi, %eax, %ebx, %ecx, %ebp, 48(%edi), MAGIC1, FUNC1)
+    F(%ecx, %edx, %esi, %eax, %ebx, %ebp, 52(%edi), MAGIC1, FUNC1)
+    F(%ebx, %ecx, %edx, %esi, %eax, %ebp, 56(%edi), MAGIC1, FUNC1)
+
+    F(%eax, %ebx, %ecx, %edx, %esi, %ebp, 60(%edi), MAGIC1, FUNC1)
+    F(%esi, %eax, %ebx, %ecx, %edx, %ebp, 64(%edi), MAGIC1, FUNC1)
+    F(%edx, %esi, %eax, %ebx, %ecx, %ebp, 68(%edi), MAGIC1, FUNC1)
+    F(%ecx, %edx, %esi, %eax, %ebx, %ebp, 72(%edi), MAGIC1, FUNC1)
+    F(%ebx, %ecx, %edx, %esi, %eax, %ebp, 76(%edi), MAGIC1, FUNC1)
+ 
+   movl 20(%esp), %ebp
+   movl %eax, 0(%ebp)
+   movl %ebx, 4(%ebp)
+   movl %ecx, 8(%ebp)
+   movl %edx, 12(%ebp)
+   movl %esi, 16(%ebp)
+   
         popl    %ebx
         popl    %esi
         popl    %edi